CN115841434A - Infrared image enhancement method for gas concentration analysis - Google Patents
Infrared image enhancement method for gas concentration analysis Download PDFInfo
- Publication number
- CN115841434A CN115841434A CN202310141646.0A CN202310141646A CN115841434A CN 115841434 A CN115841434 A CN 115841434A CN 202310141646 A CN202310141646 A CN 202310141646A CN 115841434 A CN115841434 A CN 115841434A
- Authority
- CN
- China
- Prior art keywords
- edge pixel
- pixel point
- image
- deviation
- gas
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 60
- 238000004458 analytical method Methods 0.000 title claims abstract description 22
- 238000012544 monitoring process Methods 0.000 claims abstract description 79
- 238000001914 filtration Methods 0.000 claims abstract description 39
- 238000011156 evaluation Methods 0.000 claims abstract description 15
- 238000001514 detection method Methods 0.000 claims description 20
- 238000010606 normalization Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 10
- 238000012545 processing Methods 0.000 abstract description 5
- 238000009792 diffusion process Methods 0.000 description 8
- 238000005457 optimization Methods 0.000 description 6
- 238000003708 edge detection Methods 0.000 description 5
- 238000013528 artificial neural network Methods 0.000 description 4
- 230000003044 adaptive effect Effects 0.000 description 3
- 238000012937 correction Methods 0.000 description 3
- 238000012512 characterization method Methods 0.000 description 2
- 230000002708 enhancing effect Effects 0.000 description 2
- 238000009499 grossing Methods 0.000 description 2
- 238000002372 labelling Methods 0.000 description 2
- 238000012806 monitoring device Methods 0.000 description 2
- 238000001931 thermography Methods 0.000 description 2
- 241001156002 Anthonomus pomorum Species 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000004392 development of vision Effects 0.000 description 1
- 238000013210 evaluation model Methods 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 239000000383 hazardous chemical Substances 0.000 description 1
- 238000003331 infrared imaging Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000013021 overheating Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 238000011410 subtraction method Methods 0.000 description 1
- 230000009967 tasteless effect Effects 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention relates to the technical field of image processing, in particular to an infrared image enhancement method for gas concentration analysis, which comprises the following steps: acquiring deviation edge pixel points, normal electrical equipment edge pixel points and suspected gas area images according to the infrared monitoring image; acquiring a denoised fuzzy suspected gas area image from the fuzziness index of the suspected gas area image before and after denoising; acquiring a first distance between a normal electrical equipment edge pixel point closest to the deviation edge pixel point and the deviation edge pixel point; obtaining a deviation angle set, obtaining a characteristic evaluation value of an adjacent deviation edge pixel point region according to the deviation angle set, obtaining an optimized variance according to the characteristic evaluation value and the first distance, and filtering the denoised fuzzy suspected gas region image according to the optimized variance to obtain a denoised gas leakage monitoring region image. The invention avoids the problem of poor infrared image smooth filtering effect caused by noise points near the edge of the electrical equipment.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to an infrared image enhancement method for gas concentration analysis.
Background
Along with the development of vision technique, to gas leakage's detection, the tradition utilizes gas sensor to survey gas, can only fix a point and survey, and utilizes special infrared detection instrument can carry out monitoring by a large scale to colourless tasteless gas leakage, and according to gas outgoing time and the produced ambient temperature change in the external world, carry out infrared imaging to gas leakage, detect gas leakage position and scale according to infrared thermal imaging. And gas concentration analysis is realized by detecting the obtained gas leakage gray scale characteristics and area size.
In the prior art, in the process of detecting a gas leakage image by using a thermal imager, mostly, the region position of the gas leakage is obtained by using a region segmentation and background subtraction method or a neural network model is trained by using the gas leakage image, and the profile of the leaked gas is marked by using the trained neural network to the gas leakage image. However, the method has higher requirements on image quality for both the traditional image detection method and the neural network training for gas leakage detection, when the image quality is poor and noise exists, detection errors or the model generalization capability is poor, the existing image denoising method can cause local blurring of images and influence the detection precision of tiny leakage gas, so that the concentration analysis precision of the tiny leakage gas is influenced when the gas leakage concentration is small. Based on the method, the invention provides an infrared image enhancement method for gas concentration analysis.
Disclosure of Invention
The invention provides an infrared image enhancement method for gas concentration analysis, which aims to solve the existing problems.
The invention relates to an infrared image enhancement method for gas concentration analysis, which adopts the following technical scheme:
one embodiment of the invention provides an infrared image enhancement method for gas concentration analysis, which comprises the following steps:
acquiring an infrared monitoring image, and performing deviation matching on edge pixel points on the infrared monitoring image in a gas leakage monitoring area and edge pixel points in a standard rectangular monitoring area image to obtain deviation edge pixel points, normal electrical equipment edge pixel points and a suspected gas area image;
filtering the suspected gas area image by taking a preset variance as the variance of a Gaussian filter to obtain a denoised suspected gas area image, taking the difference of the fuzziness index mean values of the suspected gas area image before and after denoising as the Gaussian denoising fuzzy degree of the suspected gas area image, and obtaining the denoised suspected gas area image according to the Gaussian denoising fuzzy degree;
acquiring the distance between the normal electrical equipment edge pixel point closest to the deviation edge pixel point and the corresponding deviation edge pixel point, and recording as a first distance;
obtaining each neighborhood range by taking each deviation edge pixel point as a neighborhood center, calculating the offset angle between all adjacent deviation edge pixel points in the neighborhood range to obtain an offset angle set, and clustering all offset angles in the offset angle set to obtain a plurality of categories; and obtaining a feature evaluation value of an adjacent deviation edge pixel point region according to the difference value of every two adjacent offset angles in each category and the number of the categories, obtaining an optimized variance according to the feature evaluation value and the first distance, and filtering the denoised fuzzy suspected gas region image by taking the optimized variance as the variance of a Gaussian filter to obtain the denoised gas leakage monitoring region image.
Preferably, the performing deviation matching on edge pixel points on the infrared monitoring image in the gas leakage monitoring area and edge pixel points in the standard rectangular monitoring area image to obtain deviation edge pixel points, normal electrical equipment edge pixel points and suspected gas area images includes:
wherein ,
the coordinates of an ith edge pixel point on an infrared monitoring image in the gas leakage monitoring area are expressed, and the device is combined with the infrared monitoring image>
Represents the coordinates of the ith edge pixel point in the standard rectangular monitoring area image, and is used for judging whether the pixel point is in the standard rectangular monitoring area image or not>
Expressing the two-dimensional coordinate Euclidean distance of the ith edge pixel point on the infrared monitoring image;
setting a deviation matching threshold M when
When the current edge pixel point is the normal edge pixel point of the electrical equipment, the ith edge pixel point is the normal edge pixel point of the electrical equipment;
when in use
Or when no edge pixel point obtained by detection exists on the standard rectangular monitoring area image, the infrared monitoring image in the gas leakage monitoring area is used as a suspected gas area image, and the ith edge pixel point is a deviation edge pixel point.
Preferably, the method for acquiring the ambiguity index of the suspected gas area image includes:
wherein ,
represents the coordinates on the image of the suspected gas area pick>
The gray value of the pixel point is judged and judged>
、/>
Indicates the width and height of the image of the area suspected of being gas, and/or>
And representing an ambiguity index of the image of the suspected gas area.
Preferably, the method for obtaining the feature evaluation value of the adjacent deviation edge pixel point region includes:
wherein ,
represents the absolute value of the offset angle difference between the jth deviation edge pixel point and the jth +1 deviation edge pixel point in the kth class, and->
Indicates that the kth class contains the total number of elements, based on the number of cells in the kth class>
Indicating the total number of categories.
Preferably, the method for obtaining the optimized variance includes:
and normalizing the feature evaluation value of the adjacent deviation edge pixel point region, and obtaining the optimized variance according to the product of twice of the normalization result and the first distance.
Preferably, the calculating offset angles between all adjacent deviation edge pixel points in the neighborhood range to obtain an offset angle set includes:
recording a deviation edge pixel point at the center of the neighborhood range as a first pixel point, acquiring a deviation edge pixel point adjacent to the first pixel point, recording the deviation edge pixel point as a second pixel point, and calculating the offset angle between the first pixel point and the second pixel point; and obtaining a third pixel point adjacent to the second pixel point, calculating the offset angle between the second pixel point and the third pixel point, and repeating the steps, wherein a set formed by the offset angles between all adjacent difference edge pixel points in the neighborhood range is marked as an offset angle set.
Preferably, the method for acquiring the standard rectangular monitoring area image comprises the following steps: and manually marking each electrical equipment monitoring area through manual work, and acquiring a standard rectangular monitoring area image according to the pixel point coordinates of the marked area.
The technical scheme of the invention at least has the following beneficial effects:
the distance between the edge point of the electrical equipment and the edge pixel point with the position deviation can reduce the Gaussian filtering fuzzy influence on the edge pixel point of the micro leakage gas, improve the contrast between the edge pixel point and the pixel point in the background area, and facilitate capturing the edge feature of the micro leakage gas with low concentration and unobvious features.
Through self characteristic index optimization of revealing gas to the distance of deviation marginal pixel, can further reduce and reveal the gas diffusion in-process, the influence of the regional characteristic is revealed to gas to the gauss filtering smoothness, and simultaneously, it is not good to also can avoid the smooth filtering effect of the noise point near electrical equipment edge.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the embodiments or the prior art descriptions will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 is a flowchart illustrating the steps of an infrared image enhancement method for gas concentration analysis according to the present invention.
Detailed Description
To further illustrate the technical means and effects of the present invention adopted to achieve the predetermined objects, the following detailed description of an infrared image enhancement method for gas concentration analysis according to the present invention with reference to the accompanying drawings and preferred embodiments shows the following detailed descriptions of the specific implementation, structure, features and effects thereof. In the following description, different "one embodiment" or "another embodiment" refers to not necessarily the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following describes a specific scheme of an infrared image enhancement method for gas concentration analysis according to the present invention in detail with reference to the accompanying drawings.
Referring to fig. 1, a flowchart illustrating a method for enhancing an infrared image for gas concentration analysis according to an embodiment of the present invention is shown.
The purpose of this example is: the noise removal effect of a local gas leakage area is improved by optimizing the existing denoising algorithm, and the detection precision of a gas leakage image is improved.
The specific scenario addressed by this embodiment is: the method mainly aims at detecting gas leakage in transformer substations, hazardous chemical plants and the like, mainly utilizes infrared thermal imaging equipment to carry out video monitoring on large-area areas, carries out denoising processing on video monitoring images, and combines the existing image denoising algorithm to realize self-adaptive optimization denoising of local areas of gas leakage. In the embodiment, a transformer substation is taken as an example for explanation, and the image denoising and enhancing process in the leakage gas concentration analysis process is described and implemented.
The method comprises the following steps:
step S101: and acquiring an infrared monitoring video image, preprocessing the video frame image, and acquiring a suspected gas area.
The method comprises the steps that a video image of an electrical equipment area is obtained through infrared monitoring equipment, in order to improve the detection efficiency of the video image, frame selection is carried out on the video image at intervals, and the infrared monitoring image is collected once every 1 second.
The collected images are segmented into regions of interest, specifically, manual labeling is carried out on each electrical equipment monitoring region through manual work, and standard rectangular monitoring region images are obtained according to pixel point coordinates of the labeling regions. Further, edge pixel point information in the gas leakage monitoring area is obtained through a Canny edge detection algorithm, and edge pixel points are used for obtaining the edge pixel point informationThe coordinate information of the monitoring area is subjected to deviation matching with edge pixel points in a standard rectangular monitoring area image, it needs to be explained that the gas leakage monitoring area is an area with the same size and the same position as the standard rectangular monitoring area on the infrared monitoring image, and the specific deviation matching method comprises the following steps: performing image two-dimensional coordinate Euclidean distance calculation through edge pixel points in the standard rectangular monitoring area image and edge pixel points in the gas leakage monitoring area image acquired in real time:
, wherein ,
represents the coordinates of the pixel point at the ith edge in the image of the gas leakage monitoring area obtained, and ` is/are `>
And expressing the coordinates of the ith edge pixel point in the standard rectangular monitoring area image. Setting a deviation matching threshold M, which is described in the embodiment with M =1 as an example, when Euclidean distance ÷ corresponding to an edge pixel is greater than or equal to>
If so, the ith edge pixel point is indicated as a normal electrical equipment edge pixel point; when the Euclidean distance of the corresponding edge pixel point is greater than or equal to>
Or when no edge pixel point obtained through detection exists on the corresponding standard rectangular monitoring area image, the current monitoring area is regarded as the existence of suspected leakage gas, the infrared monitoring image in the current monitoring area is recorded as a suspected gas area image, and the ith edge pixel point under the condition is called as an edge pixel point with position deviation, namely a deviation edge pixel point for short.
It should be noted that, in the present embodiment, the problem of missing detection of edge pixel points in the process of detecting edge pixel points by using a Canny edge detection algorithm is not considered, which is not the focus of research in the present embodiment. This embodiment mainly receives the influence to the follow-up regional monitoring accuracy that the false retrieval problem of edge pixel that the noise point caused. The edge pixel points in the standard rectangular monitoring area image are edges of the electrical equipment to be monitored in the current area, and the edge pixel points in the gas leakage monitoring area image acquired in real time comprise: and misdetection edge pixel points and gas leakage area edge pixel points caused by noise points at the edge of the electrical equipment to be monitored in the current area.
It should be noted that, because there are a plurality of electronic devices, a plurality of standard rectangular monitoring area images and a plurality of suspected gas area images are correspondingly obtained.
Step S102: and carrying out local area self-adaptive denoising according to the suspected gas area image to obtain a denoised gas leakage monitoring area image.
It should be noted that, in the video images collected by the existing infrared monitoring device, the main source of noise is caused by noise generated by a channel during transmission or channel instability generated by overheating of the infrared monitoring device. In the process of detecting gas leakage through images of suspected gas areas, the influence of image noise is easily received, generally, before image processing such as image edge detection is carried out, image denoising needs to be carried out firstly, but the existing image denoising algorithm can cause the blurring of local area texture details, the influence on tiny gas leakage detection is large, in the embodiment, the influence on the edge detection caused by noise is considered through firstly obtaining images of the suspected gas areas, the influence on the blurring of the local area texture details caused by the denoising algorithm is improved through subsequent self-adaptive denoising of the local areas, the quality of the images of the suspected gas areas is improved, and the problem of errors in the detection process of tiny gas is avoided.
The specific process of realizing the local area adaptive denoising of the suspected gas area image acquired in the step S101 is expanded as follows: a. and denoising the gas leakage monitoring area image acquired in real time through a denoising algorithm.
In the embodiment, a Gaussian denoising filter is used for denoising a gas leakage monitoring area, a Gaussian filter core for constructing 3*3 is used for filtering and denoising a suspected gas area image, and the Gaussian filter core is used for performing initial convolution on the suspected gas area imageIn the process of (1), its variance of Gaussian filter is set to
And carrying out preliminary denoising on the gas leakage monitoring area to obtain a denoised image.
In the denoised image, the fuzzy degree detection is performed between the suspected gas area images obtained in the step S101, and the Brenner gradient function is used to obtain the square of the gray difference between the suspected gas area image before denoising and the suspected gas area image after denoising, that is:
wherein ,
indicating a coordinate in the image of the suspected gas area pick>
The gray value of the pixel point at is greater than or equal to>
、/>
Indicates the width and height of an image of a suspect gas area>
The blur degree index is represented, and the larger the value is, the smaller the blur degree is.
Obtaining the Gaussian denoising fuzzy degree of the suspected gas area image through the obtained fuzzy degree index mean value difference of the suspected gas area image before and after denoising
, wherein ,/>
Represents the mean value of the ambiguity index of the images of all suspected gas areas before de-noising, based on the comparison of the values of the intensity of the detected noise and the intensity of the detected noise>
And representing the mean value of the fuzziness indexes of all the denoised images of the suspected gas areas. And obtaining the denoising effect evaluation of the suspected gas area image according to the denoised ambiguity index mean value difference. Setting a Gaussian de-noising blur degree threshold value->
When the degree of blurring is eliminated by Gaussian noise>
And if so, determining that the image blur caused by Gaussian filtering denoising exists in the current suspected gas area image, and marking as the denoised blurred suspected gas area image.
b. And acquiring a local area self-adaptive Gaussian filter variance parameter according to the de-noised fuzzy area image and the characteristics of the suspected gas leakage area.
And (b) performing self-adaptive estimation correction by using the Gaussian filtering variance on the denoised fuzzy suspected gas area image obtained in the step (a). It should be noted that the gaussian filter is a low-pass filter, and a certain loss is caused to high-frequency information during the filtering process, where the high-frequency information includes weak edge information and noise information of the micro gas. The function of the Gaussian filter variance is to control the variation difference inside the Gaussian filter kernel, and the Gaussian filter variance
The larger the smoothing degree of the pixel points in the Gaussian filter kernel is, the larger the local blurring degree is caused.
In the embodiment, in the process of Gaussian filtering and denoising, the filtering variance corresponding to a local Gaussian filtering kernel is subjected to adaptive estimation and correction through the characteristic edge position area of the electrical equipment and the edge distribution direction position of the micro gas in the denoised image and the standard monitoring area image. Specifically, the method comprises the following steps:
1. the distribution position of the edge of the electrical equipment and the distribution position of the edge pixel points with position deviation are obtained through the edge detection result of the standard monitoring area image, and the nearest edge pixel points with deviation are calculatedThe Euclidean distance of the distance between the edge pixel points of the normal electrical equipment and the corresponding deviation edge pixel points
And is denoted as a first distance. I.e. a first distance is obtained for each deviation edge pixel point correspondence.
Using Euclidean distance
In the process of carrying out deviation edge pixel point filtering on the Gaussian filtering kernel, the filtering variance is judged>
The optimization method of the self-adaptive optimization method comprises the following steps:
note that, due to the filtering variance
The size of the average value determines the allowable variance of the area around the average value, the larger the filtering variance is, the larger the Gaussian smoothness degree in the corresponding filtering kernel area is, and because the gas leakage of the electrical equipment generates an initial stage, tiny gas generally surrounds the edge of the electrical equipment, so that when the distance between the edge point of the electrical equipment and the edge pixel point with position deviation is greater>
The larger the noise pixel point is, the smaller the influence of smooth fuzzy influence caused by Gaussian filtering on the detection of the tiny gas is, 50 is a reference metric value self-set in the embodiment, the reference metric value can be adjusted according to actual detection needs, and the preset variance is based on the value of the minimum variance>
The initial Gaussian filtering variance value is adopted, and the initial Gaussian filtering is carried out by adopting a smaller filtering variance value for accurate subsequent optimization effect, so that the characteristic information of the gas leakage area is better obtained.
The Gaussian filter fuzzy influence on the edge pixel points of the tiny leakage gas can be reduced by utilizing the distance between the edge point of the electrical equipment and the edge pixel point with the position deviation, and meanwhile, the distance is reduced
The background area where the large deviation edge pixel point exists can improve the effect of Gaussian filtering smoothness, the larger the smoothness fuzzy degree of the pixel point of the background area is, the higher the neighborhood contrast of the small leakage gas edge pixel point is, and the higher the detection precision of the subsequent gas leakage area is.
2. The edge image of the standard electrical equipment is used as a Gaussian filtering variance optimization reference object, the Gaussian filtering fuzzy influence on tiny gas edge pixel points can be reduced, but the filtering of false detection edge pixel points caused by real noise pixel points can also cause the influence of the filtering effect. In the Euclidean distance
And further optimizing the Gaussian filter variance by utilizing the edge distribution characteristics of the micro gas on the basis of optimizing the Gaussian filter variance.
3. The gas leakage process can show the diffusion characteristics, namely, the distribution position of the gas leakage edge has a certain diffusion direction and a position deviation in the diffusion direction. The suspected gas leakage edge pixel points are screened according to the position offset direction between the adjacent deviation edge pixel points, and it needs to be explained that the adjacent deviation edge pixel points are not adjacent in absolute sense on the image position coordinates, but are edge pixel points with a short distance in the neighborhood range of the deviation edge pixel points, so that the problem of continuous edge discontinuity caused by weak edge characteristics of tiny gas edges on an infrared image is avoided. In the present embodiment, the size of the neighborhood range is 5*5, and the practitioner can adjust the size of the neighborhood range to be studied according to the actual situation.
4. Specifically, for any one deviation edge pixel point, all deviation edge pixel points in a 5*5 neighborhood range taking the deviation edge pixel point as a center are obtained, and the deviation edge pixel points are processed as follows:
obtaining the offset angle between the offset edge pixel point at the center and the adjacent offset edge pixel point in the 5*5 neighborhood
For the sake of convenience of study, the offset angle->
The slope between two points is used for characterization. Taking the adjacent deviation edge pixel points as new neighborhood center points, acquiring new adjacent deviation edge pixel points in the 5*5 neighborhood, and further calculating the deviation angle between the deviation edge pixel points and the adjacent deviation edge pixel points>
And repeating the above mode to obtain a deviation angle set between all adjacent deviation edge pixel points in the neighborhood of the deviation edge pixel points.
Performing mean shift clustering on all the offset angles in the offset angle set to obtain a plurality of categories, and recording the number of the obtained categories as
，/>
The larger the noise interference exists in the neighborhood range, the more the denoising is needed, the more the noise interference exists in the neighborhood range is, and the greater the denoising is needed>
The smaller the deviation angle difference between the deviation edge pixel points is, the distribution of the deviation edge pixel points accords with the diffusion characteristic of the leaked gas, and the smaller the noise removal is.
Setting the offset angles in the kth category to be arranged according to the acquired sequence
The length of the sequence is recorded as->
，/>
Sequence>
The jth offset angle of (1).
So far, the method obtains the pixel point within the 5*5 neighborhood range of any deviation edge pixel point
Each category, and the corresponding sequence of each category.
5. Using offset angle between adjacent offset edge pixels
Obtaining the offset angle difference
. By means of successive deviation angle differences >>
And (3) constructing a characteristic evaluation model according to the diffusion characteristic of the leakage gas:
wherein ,
and representing the difference of the offset angle between the jth deviation edge pixel point and the (j + 1) th deviation edge pixel point in the kth category. />
The method aims to represent the difference of the offset angles between adjacent deviation edge pixel points, so that the distribution direction of the edge pixel points has certain tendency according to the diffusion characteristics of gas, and the offset angles between the adjacent gas edge pixel points cannot be changedThere is a greater difference, which is greater>
The method aims to obtain the sum of offset angles between continuous adjacent deviation edge pixel points, and ensure that the influence degree of Gaussian filter variance on smooth blurring can be reduced in the process of simultaneously performing Gaussian filtering on a plurality of deviation edge pixel points, so that the image characteristics of the edge pixel points are highlighted.
The purpose of be in order to characterize its gaseous edge of revealing of characterization in the neighborhood scope in the quantity of distribution direction, general gaseous edge satisfies certain trend in the neighborhood scope, its diffusion direction can appear the change in the trend, but the adjacent edge in the neighborhood scope does not have the marginal pixel in a plurality of different directions almost, when having the marginal pixel in a plurality of directions, mainly receive the influence of noise pixel, combine the accurate characteristic evaluation value that obtains adjacent deviation marginal pixel regional area of marginal trend characteristic->
The method is favorable for reducing the influence of Gaussian filtering kernel smoothing and blurring on suspected gas leakage edge pixel points.
In summary, it can be known that any one of the deviation edge pixels corresponds to one feature evaluation value
And also corresponds to a first distance +>
。
It should be noted that, in the obtained categories, there may be a category with an element number of 1, and then the offset angle difference in the category is recorded as 0.
6. Estimating the distribution direction position of the deviation edge pixel points in the gas leakage monitoring area through the characteristic evaluation values of the deviation pixel points, and combining the nearest pixel point of the edge of the electrical equipment and the edge pixel point with the position deviation with the corresponding deviationEuclidean distance between edge pixels
Self-adaptive correction is carried out on the Gaussian filter variance, and the phenomenon that pixel points at the weak edge of tiny gas with large distance deviation are ignored, so that the pixel points are smoothly blurred by Gaussian filter and the details are lost is reduced. The specific gaussian filter variance is optimized as follows:
wherein ,
representing a feature evaluation value ≥ for all outlier edge pixels>
Result of the linear normalization process is taken>
Indicates that the range of Z is adjusted to->
In order to improve the Gaussian filtering effect on the deviation edge pixel points by analyzing the distribution direction position characteristics of the tiny gas edge pixel points, the judgment result is analyzed, and the judgment result is analyzed>
Represents the optimized variance of the gaussian filter kernel.
To this end, each deviation edge pixel point corresponds to one
The variance of the filter kernel corresponding to a non-deviation edge pixel point is set to ≧>
。
c. And realizing the self-adaptive denoising of the local area of the image of the gas leakage monitoring area according to the self-adaptive Gaussian filter variance parameter of the local area.
Optimizing a new Gaussian filtering process through the locally self-adaptive optimized Gaussian filtering variance obtained in the 6 th sub-step of the step S102, locally self-adaptively adjusting a fixed Gaussian filtering variance in a traditional algorithm, denoising a denoised fuzzy suspected gas area image according to the optimized Gaussian filtering kernel, obtaining a denoised gas leakage monitoring area image, and particularly setting the variance of the Gaussian filtering kernel corresponding to each pixel point in the denoising process as the optimized variance of each pixel point.
Step S103: and carrying out background difference detection on the gas leakage area according to the denoised image.
Acquiring a denoised gas leakage monitoring area image through the local adaptive denoising algorithm in the step S102, performing background difference processing on the denoised gas leakage monitoring area image and a standard rectangular monitoring area image to acquire a background-difference gas leakage area, and realizing gas identification and gas contour marking by using the existing trained gas identification neural network, wherein the marked gas contour area is the gas leakage area. The gas identification neural network can realize identification for the existing MaskRCNN network, and the input of the network is as follows: the denoised gas leakage monitoring area image is output as follows: gas contour area images. Then, the concentration analysis and estimation of the leaking gas is realized according to the area of the gas contour region and the gray average value inside the contour region, and the specific concentration analysis and estimation process is not the focus of the embodiment and is not described in detail.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (7)
1. An infrared image enhancement method for gas concentration analysis, characterized in that the method comprises the following steps:
acquiring an infrared monitoring image, and performing deviation matching on edge pixel points on the infrared monitoring image in a gas leakage monitoring area and edge pixel points in a standard rectangular monitoring area image to obtain deviation edge pixel points, normal electrical equipment edge pixel points and a suspected gas area image;
filtering the suspected gas area image by taking a preset variance as the variance of a Gaussian filter to obtain a denoised suspected gas area image, taking the difference of the fuzziness index mean values of the suspected gas area image before and after denoising as the Gaussian denoising fuzzy degree of the suspected gas area image, and obtaining the denoised suspected gas area image according to the Gaussian denoising fuzzy degree;
acquiring the distance between the normal electrical equipment edge pixel point closest to the deviation edge pixel point and the corresponding deviation edge pixel point, and recording as a first distance;
obtaining each neighborhood range by taking each deviation edge pixel point as a neighborhood center, calculating the offset angle between all adjacent deviation edge pixel points in the neighborhood range to obtain an offset angle set, and clustering all offset angles in the offset angle set to obtain a plurality of categories; and obtaining a feature evaluation value of an adjacent deviation edge pixel point region according to the difference value of every two adjacent deviation angles in each category and the number of the categories, obtaining an optimized variance according to the feature evaluation value and the first distance, and filtering the denoised fuzzy suspected gas region image by taking the optimized variance as the variance of a Gaussian filter to obtain the denoised gas leakage monitoring region image.
2. The infrared image enhancement method for gas concentration analysis according to claim 1, wherein the step of performing deviation matching on edge pixel points on the infrared monitoring image in the gas leakage monitoring area and edge pixel points in the standard rectangular monitoring area image to obtain deviation edge pixel points, normal electrical equipment edge pixel points and suspected gas area images comprises the following steps:
wherein ,
the coordinates of the ith edge pixel point on the infrared monitoring image in the gas leakage monitoring area are represented,
representing the coordinates of the ith edge pixel point in the standard rectangular monitoring area image,
expressing the Euclidean distance of two-dimensional coordinates of the ith edge pixel point on the infrared monitoring image;
setting a deviation matching threshold M when
When the current edge pixel point is the normal edge pixel point of the electrical equipment, the ith edge pixel point is the normal edge pixel point of the electrical equipment;
when in use
Or when no edge pixel point obtained by detection exists on the standard rectangular monitoring area image, the infrared monitoring image in the gas leakage monitoring area is used as a suspected gas area image, and the ith edge pixel point is a deviation edge pixel point.
3. The method of claim 1, wherein the method of obtaining the ambiguity indicator of the suspected gas area image comprises:
wherein ,
representing coordinates on the image of the area of suspected gas
The gray value of the pixel point at (a),
、
representing the suspected gas area image width and height,
indicating a ambiguity index for the image of the suspected gas region.
4. The infrared image enhancement method for gas concentration analysis according to claim 1, wherein the method for obtaining the feature evaluation value of the adjacent deviation edge pixel point region comprises:
wherein ,
represents the absolute value of the offset angle difference between the jth deviation edge pixel point and the jth +1 deviation edge pixel point in the kth class,
indicating the total number of elements contained in the kth class,
indicating the total number of categories.
5. The infrared image enhancement method for gas concentration analysis according to claim 1, wherein the obtaining method of the optimized variance comprises:
and normalizing the feature evaluation value of the adjacent deviation edge pixel point region, and obtaining the optimized variance according to the product of twice of the normalization result and the first distance.
6. The method of claim 1, wherein the computing of the offset angles between all neighboring deviation edge pixels in the neighborhood region to obtain the set of offset angles comprises:
recording a deviation edge pixel point at the center of the neighborhood range as a first pixel point, acquiring a deviation edge pixel point adjacent to the first pixel point, recording the deviation edge pixel point as a second pixel point, and calculating the offset angle between the first pixel point and the second pixel point; and obtaining a third pixel point adjacent to the second pixel point, calculating the offset angle between the second pixel point and the third pixel point, and repeating the steps, wherein a set formed by the offset angles between all adjacent difference edge pixel points in the neighborhood range is marked as an offset angle set.
7. The infrared image enhancement method for gas concentration analysis according to claim 1, wherein the standard rectangular monitoring area image is obtained by: and manually marking each electrical equipment monitoring area through manual work, and acquiring a standard rectangular monitoring area image according to the pixel point coordinates of the marked area.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310141646.0A CN115841434B (en) | 2023-02-21 | 2023-02-21 | Infrared image enhancement method for gas concentration analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310141646.0A CN115841434B (en) | 2023-02-21 | 2023-02-21 | Infrared image enhancement method for gas concentration analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115841434A true CN115841434A (en) | 2023-03-24 |
| CN115841434B CN115841434B (en) | 2023-04-21 |
Family
ID=85580005
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310141646.0A Active CN115841434B (en) | 2023-02-21 | 2023-02-21 | Infrared image enhancement method for gas concentration analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115841434B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116245880A (en) * | 2023-05-09 | 2023-06-09 | 深圳市银河通信科技有限公司 | Electric vehicle charging pile fire risk detection method based on infrared identification |
| CN116311079A (en) * | 2023-05-12 | 2023-06-23 | 探长信息技术(苏州)有限公司 | Civil security engineering monitoring method based on computer vision |
| CN116311383A (en) * | 2023-05-16 | 2023-06-23 | 成都航空职业技术学院 | Intelligent building power consumption management system based on image processing |
| CN116563799A (en) * | 2023-07-11 | 2023-08-08 | 山东昆仲信息科技有限公司 | Dust online monitoring method based on video monitoring |
| CN116844116A (en) * | 2023-09-01 | 2023-10-03 | 山东乐普矿用设备股份有限公司 | Underground comprehensive safety monitoring system based on illumination control system |
| CN117470450A (en) * | 2023-12-27 | 2024-01-30 | 广州煌之牌包装机械有限公司 | Method and system for detecting leakage of multi-head type rolling cover of penicillin bottle |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107451590A (en) * | 2017-07-19 | 2017-12-08 | 哈尔滨工程大学 | Gas detection identification and concentration method for expressing based on EO-1 hyperion infrared image |
| US20190003919A1 (en) * | 2015-12-15 | 2019-01-03 | Konica Minolta, Inc. | Image processing device for gas detection, image processing method for gas detection, image processing program for gas detection, computer-readable recording medium having image processing program for gas detection recorded thereon, and gas detection system |
| CN109727202A (en) * | 2017-10-27 | 2019-05-07 | 清华大学 | Gas infrared video Enhancement Method, device and storage medium |
| CN112763155A (en) * | 2021-02-02 | 2021-05-07 | 南京信息工程大学 | Pressure container gas leakage detection method |
| CN113720544A (en) * | 2021-08-16 | 2021-11-30 | 天津大学 | Weak sound wave or tiny gas leakage detection system and method |
| CN114596329A (en) * | 2020-12-04 | 2022-06-07 | 中国科学院深圳先进技术研究院 | Gas image enhancement and gas leakage detection method and system |
-
2023
- 2023-02-21 CN CN202310141646.0A patent/CN115841434B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190003919A1 (en) * | 2015-12-15 | 2019-01-03 | Konica Minolta, Inc. | Image processing device for gas detection, image processing method for gas detection, image processing program for gas detection, computer-readable recording medium having image processing program for gas detection recorded thereon, and gas detection system |
| CN107451590A (en) * | 2017-07-19 | 2017-12-08 | 哈尔滨工程大学 | Gas detection identification and concentration method for expressing based on EO-1 hyperion infrared image |
| CN109727202A (en) * | 2017-10-27 | 2019-05-07 | 清华大学 | Gas infrared video Enhancement Method, device and storage medium |
| CN114596329A (en) * | 2020-12-04 | 2022-06-07 | 中国科学院深圳先进技术研究院 | Gas image enhancement and gas leakage detection method and system |
| CN112763155A (en) * | 2021-02-02 | 2021-05-07 | 南京信息工程大学 | Pressure container gas leakage detection method |
| CN113720544A (en) * | 2021-08-16 | 2021-11-30 | 天津大学 | Weak sound wave or tiny gas leakage detection system and method |
Non-Patent Citations (2)
| Title |
|---|
| 刘路民根 等: "一种基于形状的红外图像泄漏气体检测方法" * |
| *** 等: "基于红外成像的乙烯气体泄漏检测" * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116245880A (en) * | 2023-05-09 | 2023-06-09 | 深圳市银河通信科技有限公司 | Electric vehicle charging pile fire risk detection method based on infrared identification |
| CN116245880B (en) * | 2023-05-09 | 2023-07-18 | 深圳市银河通信科技有限公司 | Electric vehicle charging pile fire risk detection method based on infrared identification |
| CN116311079A (en) * | 2023-05-12 | 2023-06-23 | 探长信息技术(苏州)有限公司 | Civil security engineering monitoring method based on computer vision |
| CN116311079B (en) * | 2023-05-12 | 2023-09-01 | 探长信息技术(苏州)有限公司 | Civil security engineering monitoring method based on computer vision |
| CN116311383A (en) * | 2023-05-16 | 2023-06-23 | 成都航空职业技术学院 | Intelligent building power consumption management system based on image processing |
| CN116311383B (en) * | 2023-05-16 | 2023-07-25 | 成都航空职业技术学院 | Intelligent building power consumption management system based on image processing |
| CN116563799A (en) * | 2023-07-11 | 2023-08-08 | 山东昆仲信息科技有限公司 | Dust online monitoring method based on video monitoring |
| CN116563799B (en) * | 2023-07-11 | 2023-09-22 | 山东昆仲信息科技有限公司 | Dust online monitoring method based on video monitoring |
| CN116844116A (en) * | 2023-09-01 | 2023-10-03 | 山东乐普矿用设备股份有限公司 | Underground comprehensive safety monitoring system based on illumination control system |
| CN116844116B (en) * | 2023-09-01 | 2023-12-05 | 山东乐普矿用设备股份有限公司 | Underground comprehensive safety monitoring system based on illumination control system |
| CN117470450A (en) * | 2023-12-27 | 2024-01-30 | 广州煌之牌包装机械有限公司 | Method and system for detecting leakage of multi-head type rolling cover of penicillin bottle |
| CN117470450B (en) * | 2023-12-27 | 2024-03-01 | 广州煌之牌包装机械有限公司 | Method and system for detecting leakage of multi-head type rolling cover of penicillin bottle |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115841434B (en) | 2023-04-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115841434A (en) | Infrared image enhancement method for gas concentration analysis | |
| CN109743879B (en) | Underground pipe gallery leakage detection method based on dynamic infrared thermography processing | |
| CN108921176B (en) | Pointer instrument positioning and identifying method based on machine vision | |
| CN111325721A (en) | Gas leakage detection method and system based on infrared thermal imaging | |
| CN116188462B (en) | Noble metal quality detection method and system based on visual identification | |
| CN115775250B (en) | Golden finger circuit board defect rapid detection system based on digital image analysis | |
| CN116721107B (en) | Intelligent monitoring system for cable production quality | |
| CN111598869B (en) | Method, equipment and storage medium for detecting Mura of display screen | |
| CN115018844A (en) | Plastic film quality evaluation method based on artificial intelligence | |
| CN116228780B (en) | Silicon wafer defect detection method and system based on computer vision | |
| CN111738931B (en) | Shadow removal algorithm for aerial image of photovoltaic array unmanned aerial vehicle | |
| CN116958125B (en) | Electronic contest host power supply element defect visual detection method based on image processing | |
| CN113155839A (en) | Steel plate outer surface defect online detection method based on machine vision | |
| CN116168037A (en) | Method and system for calculating bending degree of wire crimping based on image processing | |
| CN114332081B (en) | Textile surface abnormity determination method based on image processing | |
| CN117764864B (en) | Nuclear magnetic resonance tumor visual detection method based on image denoising | |
| CN115029704A (en) | Intelligent control method and device for stainless steel pickling process | |
| CN110186929A (en) | A kind of real-time product defect localization method | |
| CN113052794A (en) | Image definition recognition method based on edge features | |
| CN114742849B (en) | Leveling instrument distance measuring method based on image enhancement | |
| CN114998346B (en) | Waterproof cloth quality data processing and identifying method | |
| CN116188510A (en) | Enterprise emission data acquisition system based on multiple sensors | |
| CN113554688B (en) | O-shaped sealing ring size measurement method based on monocular vision | |
| CN115761606A (en) | Box electric energy meter identification method and device based on image processing | |
| CN114742832A (en) | Welding defect detection method for MiniLED thin plate |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |